Prior to the present invention, the advent of increasingly efficient and sensitive catalysts for the polymerization of propylene has caused the polypropylene industry to recognize the importance of the control of various trace impurities in the propylene feedstock. Carbonyl sulfide has been found to be one of the most troublesome impurities, causing catalyst carryover, increased ash content of product, the production of undesirable large quantities of actactic by-product, and gross inefficiencies in catalyst life and conversion rates. These difficulties tend to be more pronounced and important in proportion to the increase in yield or efficiency otherwise observed in new catalysts.
COS levels in some polypropylene feedstocks may range from an acceptable level of .ltoreq.50 ppb (parts per billion by weight) to totally unacceptable levels of over 2 ppm (parts per million by weight). In the past, a solid NaOH bed has been used commercially but NaOH alone is not capable commercially of lowering the COS concentration to .ltoreq.50 ppb.
Concentrations of COS in the range of a few parts per million (e.g., 1-10 ppm) are very difficult to separate from C.sub.3 H.sub.6 by fractional distillation because the boiling point of COS differs from C.sub.3 H.sub.6 by only 3.4.degree. C. Also, COS is not completely removed from propylene by the usual sulfur-removal processes such as caustic scrubbing or amine-type scrubbing due primarily to the slow rate of hydrolysis of COS.
While it is known to hydrolyze COS over a platinum sulfide/alumina catalyst, such catalysts have not been used for the hydrolysis of small amounts of COS in propylene, possibly because of the fear of deposition of polymer on the surface of the catalyst, and the difficulty of regenerating the catalyst if such a deposition occurs.